Background: The Bispectral Index (BIS) and the Patient State Index (PSI) quantify depth of anesthesia by analyzing the electroencephalogram. The authors examined the response of BIS and PSI to sevoflurane anesthesia.
Methods: In 22 patients, sevoflurane anesthesia was induced by inhalation with a tight-fitting facemask and was maintained via a laryngeal mask. Sevoflurane concentrations were increased until burst suppression occurred and subsequently decreased until BIS recovered to values above 60. This procedure was repeated twice until patients underwent intubation for subsequent surgery. End-tidal sevoflurane concentrations, BIS, and PSI were recorded simultaneously. The performance of PSI and BIS to predict the estimated sevoflurane effect site concentration, as derived from simultaneous pharmacokinetic and pharmacodynamic modeling, was compared by determination coefficients (rho(2)) and prediction probabilities (P(K)).
Results: A significant (P < 0.001) correlation between BIS and PSI was found (r(2) = 0.75), and a close sigmoid relation between sevoflurane effect site concentration and both BIS (rho(2) = 0.84 +/- 0.09) and PSI (rho(2) = 0.85 +/- 0.15) was observed. The maximum sevoflurane electroencephalographic effect resulted in PSI values (1.3 +/- 4.3) that were significantly (P = 0.019) lower than BIS values (7.9 +/- 12.1), and the effect site efflux constant k(e0) was significantly smaller (P = 0.001) for PSI (0.13 +/- 0.08 min(-1)) than for BIS (0.24 +/- 0.15 min(-1)). The probability of BIS (P(K) = 0.80 +/- 0.11) to predict sevoflurane effect site concentration did not differ (P = 0.76) from that of PSI (P(K) = 0.79 +/- 0.09).
Conclusions: The BIS reacted faster to changes in sevoflurane concentrations, whereas the PSI made better use of the predefined index range. However, despite major differences in their algorithms and minor differences in their dose-response relations, both PSI and BIS predicted depth of sevoflurane anesthesia equally well.